The real problem with 8.f4 is the move 8...Bg4! which gives Black a good game.
8.Bb5 Qb4+ seems nonsensical to me: isn't white just a pawn up for nothing after 9.Qxb4 Bxb4+ 10.c3?
A better try is unpinning the c6 knight first by 8...Bd7 9.Nc3 Qb4 (now this makes much more sense 10.Qf4! (resembles computer play, but it's actually a very strong move) when IMO white is better.
There was an article about the Yaroslavtsev variation (3...Nxe4) in an issue of "Secrets of Opening Surprises"- can't recall which issue right now.
Had a game in the Petrov, where I had a slight advantage but couldn't really convert it into anything much and let it slip away.
The computer labels 8. f4 as a mistake (does weaken the kingside but I think the king is safe enough), but the games I looked at where white simply develops the pawn is lost. I thought this move would aim to hold the pawn without tying pieces down to its defence.
Is trying to keep the pawn too much of a hinderance, I should just try for a positional advantage?
Computer analysis: